Technical Summary: Hemoglobin-to-Red Blood Cell Distribution Width Ratio and All-Cause Mortality in Older Adults

1. Problem Statement

The global aging population faces increasing risks of cognitive decline and mortality, necessitating the identification of modifiable risk factors and accessible biomarkers for early stratification. While the Hemoglobin-to-Red Blood Cell Distribution Width Ratio (HRR)—an integrative index reflecting oxygen-carrying capacity and erythrocyte volume heterogeneity—has been linked to adverse health outcomes, the mechanistic pathways connecting HRR to mortality remain unclear. Specifically, it is unknown whether cognitive impairment mediates this relationship and which specific cognitive domains (e.g., processing speed vs. episodic memory) drive this association. This study addresses the gap in understanding the causal pathways linking HRR to all-cause mortality in older adults and quantifies the mediating role of cognitive function.

2. Methodology

Study Design and Data Source:
This was a prospective cohort study utilizing data from the National Health and Nutrition Examination Survey (NHANES) 2011–2014 cycles. The study followed 2,433 community-dwelling adults aged ≥60 years, with mortality outcomes tracked via the National Death Index (NDI) through December 31, 2019. The analysis accounted for the complex multistage sampling design of NHANES using survey weights, stratification, and clustering variables.

Exposure:
The primary exposure was the Hemoglobin-to-Red Blood Cell Distribution Width Ratio (HRR), calculated as Hemoglobin (g/dL) divided by RDW (%). Participants were categorized into quartiles (Q1: <0.93 to Q4: >1.17) based on the study population distribution.

Mediator:
Cognitive function was assessed using four validated neuropsychological tests:

• Digit Symbol Substitution Test (DSST): Processing speed and executive function.
• Animal Fluency Test (AFT): Executive function and semantic fluency.
• CERAD Immediate and Delayed Recall Tests: Episodic memory.
  A global composite Z-score was calculated, and cognitive impairment was defined as a Z-score < −1 standard deviation (SD).

Outcome:
The primary outcome was all-cause mortality.

Statistical Analysis:

• Associations: Multivariable linear regression (for cognitive outcomes) and Cox proportional hazards models (for mortality) were used, adjusting for demographics, lifestyle factors, and chronic conditions (Model 4).
• Mediation: Causal mediation analysis under the counterfactual framework was employed to quantify the indirect effect of cognitive impairment on the HRR-mortality relationship. Both binary (impairment status) and continuous (Z-score) mediators were tested.
• Domain-Specific Analysis: Separate mediation models were run for DSST, AFT, CERAD tests, and the global Z-score to identify dominant pathways.
• Robustness: Competing risk models (Fine-Gray) were used to account for dementia-related deaths. Sensitivity analyses included excluding early deaths (2 and 5 years), Inverse Probability of Treatment Weighting (IPTW), and E-value assessment to evaluate unmeasured confounding.

3. Key Results

Baseline Characteristics:
Participants in the highest HRR quartile (Q4) were significantly younger, more likely to be male, had lower BMI, and demonstrated superior cognitive performance (higher DSST, CERAD-DRT, and global Z-scores) compared to the lowest quartile (Q1).

HRR and Mortality:
A significant dose-response relationship was observed between HRR and all-cause mortality. In the fully adjusted model, the highest HRR quartile (Q4) was associated with a 53% reduction in mortality risk compared to the lowest quartile (Q1) (HR = 0.47, 95% CI: 0.36–0.63, P for trend < 0.001).

HRR and Cognitive Function:
Higher HRR was positively associated with better performance in processing speed (DSST), executive function (AFT), and the global cognitive Z-score in fully adjusted models. No significant association was found with immediate episodic memory (CERAD-IRT).

Mediation Analysis:
Cognitive impairment significantly mediated the association between HRR and all-cause mortality.

• Overall Mediation: The indirect effect accounted for approximately 19%–21% of the total effect (Binary mediator: 18.78%; Continuous Z-score: 21.25%).
• Domain Specificity: Processing speed (DSST) was identified as the dominant cognitive pathway, mediating 25.74% of the association. Global cognition mediated 17.43%, and executive function (AFT) mediated 5.84%. In contrast, episodic memory showed negligible mediation (0.46%, P = 0.844).

Robustness:
Findings remained consistent across competing risk models and sensitivity analyses. The E-value for the total effect was 3.42, indicating that an unmeasured confounder would need a risk ratio of at least 3.42 with both exposure and outcome to fully explain the observed association.

4. Key Contributions

• Quantification of Mediation: This is the first study to quantify the mediating role of cognitive impairment in the HRR-mortality pathway, establishing that cognitive decline accounts for roughly one-fifth of the risk.
• Pathway Differentiation: The study distinguishes between cognitive domains, revealing that HRR influences mortality primarily through processing speed (reflecting cerebrovascular and white matter integrity) rather than episodic memory (reflecting hippocampal/neurodegenerative pathology).
• Biomarker Utility: It validates HRR, a routine and cost-effective hematological parameter, as a potential screening tool for risk stratification in aging populations.

5. Significance and Claims

The authors claim that their findings provide evidence for a "hematology–cognition–mortality" pathway. The study suggests that HRR influences mortality risk primarily through cerebrovascular-small vessel disease mechanisms (indicated by the dominance of processing speed mediation) rather than classical neurodegenerative pathways (indicated by the lack of episodic memory mediation).

The significance of the work lies in its potential for public health application:

1. Risk Stratification: HRR can serve as a cost-effective, routinely available biomarker in primary care to identify older adults at elevated risk for cognitive decline and mortality.
2. Targeted Intervention: By identifying the specific pathway (cerebrovascular/white matter), the study supports interventions targeting reversible causes of low HRR (e.g., anemia, inflammation, nutritional deficiencies) to potentially mitigate mortality risk.
3. Screening Protocols: The dominance of processing speed suggests that brief cognitive screening tools like the DSST could be integrated into community health protocols to detect vascular cognitive impairment early.

The authors maintain a modest stance, acknowledging that while causal mediation analysis and sensitivity tests support robustness, the observational design cannot definitively prove causality, and longitudinal trajectory data or neuroimaging validation are needed for future confirmation.